In existing display apparatuses such as LED (Light Emitting Diode) displays, OLED (Organic Light Emitting Diode) displays, PDP (Plasma Display Panel) displays and LCDs (Liquid Crystal Displays), multiple pixels arranged in matrix are usually disposed, wherein each pixel comprises three or four sub-pixels of different colors. Based on such a structure, each sub-pixel is input with a corresponding gray scale signal in one frame such that the pixel can present a certain color to form the displayed image. It can be seen that in the above display apparatuses, the size and pitch of the pixels determine the resolution of the displayed image. However, emergence of high resolution algorithm breaks through the limitation on the image resolution by the physical resolution of the pixels.
A new image processing approach is high resolution algorithm, through which a relatively low physical resolution can be raised to a relatively high virtual resolution for a certain sub-pixel arrangement by taking advantage of the characteristics of human eyes' spatial resolution and in a way such as sub-pixel sharing, whereby not only an optimized display effect but also advantages such as low power consumption and low processing difficulty can be achieved. For example, through a high resolution algorithm, a gray scale value of a sub-pixel at every position can be obtained by appropriately processing the image to be displayed, whereby a high resolution display picture subjected to an image conversion can be obtained.
In existing high resolution algorithms, there is a step for specially processing particular patterns in the image to be displayed. All particular patterns in the image are required to be processed effectively, meanwhile, the relationship between a general processing and a special processing performed on each real pixel needs to be considered, and also the processing of some real pixels also requires referring to previous processing results. Therefore, in a case where algorithm efficiency is considered, it becomes a problem to be solved in the art how to complete an image conversion flow integrated with particular pattern processing in one traversal without repetition or missing.